Professional Documents
Culture Documents
a r t i c l e i n f o a b s t r a c t
Article history: Biot numbers (Bi) relate internal and external resistances to mass or heat fluxes and are useful to
Received 23 February 2010 identify controlling mechanisms. Their variation during drying received little attention despite its
Received in revised form 25 June 2010 importance in model selection. Therefore, variable Bi were analysed in drying of wheat, a low-mois-
Accepted 4 July 2010
ture product, and in the formation of a low-calorie, dehydrated apple-leather (LCAL), a high moisture
Available online 27 July 2010
product. Wheat drying was predicted with a numerical solution from a m.c. of 0.35 kg water/kg dry
matter, to find a variable mass transfer Bi that shifted the controlling mechanism from internal–exter-
Keywords:
nal to internal. From a lower m.c. of 0.205 kg water/kg dry matter, the entire process was internally
Biot number
Drying
controlled. Drying of shrinking LCAL was solved with the energy balance. The mass transfer Bi varied
Food from 0.5 to 1600, while the heat transfer Bi, only from 0.25 to 0.5, and a modified version, from 0.25
to 0.
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction drying is that the mass transfer Biot number may vary during the
process (Giner and Mascheroni, 2001). Another problem that usu-
The Biot number is a dimensionless group that relates a mea- ally appears in drying is that moisture content in solids is not an
sure of the rate of internal heat conduction or mass diffusion in sol- uniform field across the solid–fluid interface unlike the vapour
ids with a measure of the rate of external convection. It appears in pressure, so a sorption relationship must be used to transform
partial differential equations of transport phenomena, when the moisture contents at the solid surface into vapour pressure (Crap-
surface boundary condition relating internal and external fluxes iste and Rotstein, 1997).
is expressed in dimensionless form. Adhikari et al. (2005), during a research aimed at predicting
A typical graph used by many chemical and food engineers was stickiness during spray drying of a sugar-rich solution, have pre-
published first by Gurney and Lurie (1923), who analytically solved dicted the evolution of Biot numbers during the process in a
the unsteady state heat conduction equation with an infinite series, 120 lm initial diameter drop. They found internal control for
providing the predictions of dimensionless local temperature in- the mass transfer rate from the start, and a predominantly exter-
side the solids (the dependent variable), as a function of the dimen- nal control for heat transfer. However, additional contributions
sionless time for several dimensionless positions (the independent are required to study the definition and variation of the heat
variables) and several Biot numbers. An inherent assumption is and mass transfer Biot numbers in larger food particles in order
that the Biot number remains constant during the process. The to assess changes in the controlling mechanisms. This feature is
analytical series eventually converges in one term at long times, in itself relevant, and, besides, provides better background for
allowing straight lines to be plotted in a semilog graph. These selection of suitable mathematical models and methods of solu-
graphs were and are still used to solve heat transfer problems for tion. This is especially important nowadays, where much of the
regular geometries as slab, infinite cylinder and sphere, as well contemporary drying research resorts to empirical kinetic mod-
as three-dimensional regular geometries (Singh and Heldman, els, which do not explain the drying mechanisms (Akpinar et
1993). al., 2003).
The risk of assuming the validity of the heat and mass transfer Therefore, the objectives of this work were (1) to calculate mass
analogy for using these graphs to solve mass transfer processes as and heat transfer Biot numbers during drying of foods and (2) to
analyse their behaviours.
To this end two case studies are presented: wheat drying
* Corresponding author at: Centro de Investigación y Desarrollo en Criotecnología kinetics, using data from a previous study, and formation of a
de Alimentos (CIDCA), UNLP-CONICET, Calle 47 y 116 (B1900 AJJ) La Plata,
low-calorie, apple-leather during hot-air dehydration, a novel
Argentina. Tel.: +54 221 4249287; fax: +54 221 42890741.
E-mail address: saginer@ing.unlp.edu.ar (S.A. Giner). product.
0260-8774/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jfoodeng.2010.07.005
S.A. Giner et al. / Journal of Food Engineering 101 (2010) 214–222 215
Nomenclature
A transfer area of food in Eq. (16), m2 nt time index for numerical integration, t = (nt 1) Dt
aw water activity, decimal Nmh parameter in Eq. (3)
Bic heat transfer Biot number, dimensionless ps saturation vapour pressure of pure water, Pa
Bicd heat transfer Biot number, defined in Eq. (19), dimen- pv vapour pressure at food surface, Pa
sionless pva partial vapour pressure in the air bulk, Pa
Bim mass transfer Biot number, dimensionless pvad dimensionless vapour pressure at food surface
Cmh parameter in Eq. (3) pv0 initial vapour pressure at food surface, Pa
Cp specific heat of food defined in Eq. (17), J kg1 dry mat- r radial coordinate in a wheat grain, m
ter K1 R equivalent spherical radius in wheat, m
cw local concentration of water in food in Eq. (13), kg m3 rad dimensionless radial coordinate in a wheat grain
cw0 initial local concentration of water in food, kg m3 Rg gas constant, J mol1 K1
Deff effective diffusion coefficient of water in the food, t time, s
m2 s1 T grain temperature in Eq. (3), °C
Deff0 effective diffusion coefficient of water in food at W0 de- Ta air temperature, °C
fined in Eq. (20), m2 s1 Tm time-dependent average temperature of food in Eq. (16),
D1 Arrhenius preexponential factor in Eq. (22), m2 s1 °C
D11 parameter in Eq. (4) Tm0 initial food temperature, °C
D12 parameter in Eq. (4) x coordinate along food thickness in Eq. (13), m
Ea activation energy in food drying, J mol1 W local moisture content of food, kg water/kg dry matter
hT air-food heat transfer coefficient, W m2 K1 Wad dimensionless moisture content of food
hTd air-food heat transfer coefficient, defined in Eq. (19), We equilibrium moisture content, kg water/kg dry matter
W m2 K1 W0 initial moisture content of food, kg water/kg dry matter
i space index for numerical integration of Eq. (13), x = (i
1) Dx Greek symbols
I total number of points along x for numerical integration DHws heat of desorption of water from food in Eq. (16), J kg1
of Eq. (13) Dx distance step for numerical integration of Eq. (13), m
Kmh parameter in Eq. (3) Dt time step for numerical integration of Eq. (13), s
kp air-food mass transfer coefficient, q density of food at moisture content W, kg m3
kg vapour m2 s1 Pa1 qd density of food dry matter in Eq. (19), kg m3
kT food thermal conductivity, W m1 K1 qd0 ratio of the dry matter in the food to its initial volume,
L variable thickness of food during drying, m kg m3
L0 initial food thickness in Eq. (19), m q0 initial density of food, kg m3
ms food dry mass in Eq. (16), kg qw density of liquid water, kg m3
mwev rate of water evaporation in food in Eq. (16), kg s1
n parameter in Eq. (20)
2. Mathematical modelling On the other hand, water activity aw = pv/ps and W, the local
moisture content in the grain (kg water/kg dry matter), are related
2.1. Model describing wheat drying kinetics
@W
r¼0 ¼0 t>0 ð2:2Þ
@r
@W
r¼R qd0 Deff ¼ kp ðpv pva Þ t>0 ð2:3Þ
@r r¼R
where qd0 is the ratio of the dry mass to the grain volume, qd0 = q0/ Fig. 1. Experimental (d) and predicted (—) data of wheat grain density (q) as a
(1 + W0), q0 being the grain particle density at the initial moisture function of moisture content (W) in kg water/kg dry matter. The predictive
expression used was an empirical third-grade polynomial, q = 23463.5
content, W0. Wheat density data measured by Giner and Denisienia W3 11872.3 W2 + 1370.8 W + 1310.0. The coefficient of determination, r2 was of
(1996) was analysed here to develop a third-grade polynomial cor- 0.999. The correlation is valid for moisture contents W between 0 and 0.3 kg water/
relation with moisture content (Fig. 1). kg dry matter.
216 S.A. Giner et al. / Journal of Food Engineering 101 (2010) 214–222
by the sorption isotherm. Symbol pv is the water vapour pressure The formal definition of the mass transfer Biot becomes,
at the food surface, while ps is the saturation vapour pressure of
k R
pure water, predicted with a correlation published earlier (Giner Bim ¼ h p i ð9Þ
1
and Mascheroni, 2001). The symbol pva is the partial pressure of ps
@W
@aw
qd0 Deff
r¼R
water vapour in the air bulk and kp is the mass transfer coefficient
in kg water vapour/m2 s Pa. The modified Henderson equation was This dimensionless number expresses the ratio of the internal
selected for the aw–W relationship, which has been widely used to resistance to mass transfer to the external one. The internal is rep-
predict drying processes in grains. Considering that aw = pv/ps. resented by the grain equivalent radius (R) and the reciprocal of
2 31=Nmh the diffusion coefficient (Deff) while the external, by the reciprocal
Ln 1 ppv of the mass transfer coefficient (kp).
W ¼ 0:014 5
s
ð3Þ By using Eq. (3), the following derivative is found:
K mh ðT þ C mh Þ
1
For breadmaking wheat, Kmh, Cmh and Nmh are 2.301 105, @W 1 1 N mh 1
¼ 0:01 1 ½Lnð1 aw Þ
55.815 and 2.2857, respectively (Brooker et al., 1992). Both in @aw N mh K mh ðT þ C mh Þ 1 aw
wheat (Giner and Mascheroni, 2001) and soybean drying (Gely ð10Þ
and Giner, 2007), no quantitative differences were observed in By replacing the expressions of Eq. (7)–(9) into Eq. (6), the def-
the predicted mean grain moisture content as a function of time, inition of the mass transfer Biot number, Bim becomes.
between the isothermal model of Eqs. (1)–(3) and the non-isother-
kp R
mal model that adds heat transfer coupling. The enormous differ- Bim ¼ 2 3
1
ences between heat and mass transfer rate, the small size of 14 N mh
ps
0:01 N1mh 1 K mh ðTþC
1
mh Þ
½Lnð1 aw Þ 1aw 5
1
qd0 Deff
grains and its low-moisture nature were found to be the causes
r¼R
of this nominally isothermal drying. Therefore, only the mass
ð11Þ
transfer model was solved for wheat. All properties used in wheat
were taken from Giner and Mascheroni (2001), except the diffusion In a previous work, Parti (1993) had defined Bim as ky R/(qDeff),
coefficient which was developed by the same authors in a subse- where ky is the mass transfer coefficient for a flux equation that
quent article (Giner and Mascheroni, 2002). utilises the driving force in air humidity units. While this variable
is directly related with the partial pressure of vapour in air, the
Ea
Deff ¼ ½D11 þ D12 ðW 0 0:1891Þ exp ð4Þ expression of Parti does not include the conversion between food
Rg ðT a þ 273:16Þ
moisture content and air humidity or vapour pressure at the inter-
where D11 = 5.046 107 m2 s1 and D12 = 54.44 107 m2 s1/ face and is therefore incorrect.
(kg water/kg dry matter). The value of the activation energy was On the other hand, if Biot numbers are predicted for non-iso-
Ea = 27,184 J mol1. The symbol Ta is the drying air temperature, thermal systems and/or using moisture-dependent diffusion coef-
in °C. ficients, the value of Deff must be evaluated at the temperature
and/or local moisture content prevailing at the surface at any time.
2.1.1. Calculation of the mass transfer Biot number Besides, models predicting shrinkage during drying, must update
To establish an uniform field variable across the grain–air inter- the product characteristic dimension in each time step of the
face, the surface boundary condition Eq. (2.3) was expressed in an numerical calculation. Therefore, apart from the variable nature
equivalent, alternative form, by including the following identity. of the conversion factor (oW/oaw), possible variations in Deff and
R mean that the mass transfer Biot number would change consid-
@W @W @pv
ð5Þ erably during drying.
@r @pv @r
which, once included into the surface boundary condition (Eq. 2.1.2. Conventional heat transfer Biot number
(2.3.), becomes The heat transfer Biot number.
!
@pvs kp hT R
r¼R qd0 Deff ¼ ðpv pva Þ t>0 ð6Þ Bic ¼ ð12Þ
@r @W kT
@pv
where qd0 is the dry matter divided by the grain volume. To find the was calculated at the same conditions as for the mass transfer coun-
mass transfer Biot number, variables must be expressed in dimen- terpart. As hT, the interfacial heat transfer coefficient (Giner and
sionless form Mascheroni, 2001) have a weak temperature dependence, and as
kT is a smooth function of moisture content (Sokhansanj and Bruce,
W We r pv pva 1987), Bic is not expected to vary considerably during drying.
W ad ¼ ; r ad ¼ ; pvad ¼ ð7Þ
W0 We R pv0 pva
where R, is the grain equivalent radius, was taken as 2.0 103 and 2.2. Drying model for the formation of a low-calorie apple-leather
1.8 103 m for initial moisture contents (W0) of 0.35 and 0.205 kg (LCAL)
water/kg dry matter, respectively (Giner and Mascheroni, 2002).
The symbol We stands for the equilibrium moisture (kg water/kg 2.2.1. Product characteristics
dry matter), calculated with Eq. (3) for the temperature and relative The LCAL was formulated to experience gelation during drying
humidity of the drying air. Knowing that pv is the water vapour by the saccharide-acid-high methoxyl pectin mechanism. A previ-
pressure at the solid surface (r = R), pv0 represents its initial value. ous formulation (Leiva Díaz et al., 2009) composed of peeled apple
Vapour pressure drops from pv to pva in the air film surrounding puree, sucrose and citric acid, was modified by replacing sucrose
the grain. The derivative of moisture with respect to vapour by polydextrose which is a polysaccharide and thickener that pro-
pressure can be more easily evaluated from the sorption isotherm vides only a quarter of the energy intake of sucrose (Cicuttin and
model if we consider the following identity Giner, 2006). The initial formulation had the following composition
(% w/w): water, 80 (or 4 kg water/kg dry mater); carbohydrates,
@W 1 @W 8.86; proteins, 0.27; lipids, 0.087; citric acid, 0.17; polydextrose,
ð8Þ
@pv ps @aw 10; sucralose, 0.02. It was dried in a forced convection oven with
S.A. Giner et al. / Journal of Food Engineering 101 (2010) 214–222 217
dT m
ms C p ¼ hT AðT a T m Þ DHws mwev ð16Þ
dt
where Tm is the average product temperature and °C, Ta the drying
air temperature. The value of DHws, the heat of desorption of water
from the food, was considered the same as that for pure water in the
moisture range considered, as observed in sugar-rich products by
Kaymak-Ertekin and Gedik (2004). In turn, Cp is the variable specific
heat based on the product dry mass, (J kg1 dry matter °C), whose
expression was developed by rearranging the well-known Siebel
correlation for high moisture foods (Mohsenin, 1980)
where D1 = 7.547 107 m2 s1 is the preexponential factor, while 2.2.6. Calculation of the mass transfer Biot number
Ea, the activation energy, was found to be 19,715 J mol1. The factor By using the formal definition established in Eq. (11) and adapt-
enclosed by square brackets is Deff0, while the multiplier written in ing it to this case study where variable diffusion coefficient, and
curly brackets, represents the ratio of L/L0. The effective diffusion variable leather thickness are predicted as a function of time, a spe-
coefficient increases for increasing water concentration cw and cific definition of the mass transfer Biot number for shrinking
product temperature, Tm. The correlation of Eq. (22) was developed materials was reached using Eqs. (14.3), (23) and (24), which
by Cicuttin and Giner (2006) and follows structural collapse con- was computed as follows:
cepts of a products in the rubbery state (Aguilera, 2005). Symbol PI1
q0 stands for the initial product density = 1058.5 kg m3. To convert kp i¼1 Dxi;nt
Bim ¼ ( ) ð28Þ
from moisture content on dry matter basis to water concentration
1 @W I;nt DI;nt
in kg m3, Eq. (15) was used in the form ps @aw
qd q 2
ð1þqwd W Þ
W
cw ¼ q ð23Þ The derivative of moisture content at the surface with respect to
1þW
water activity was calculated in this LCAL with a Halsey equation,
Concerning the dependence of total product density with mois- whose parameters were previously fitted to experimental data
ture content, the model developed earlier by Leiva Díaz et al. measured by an Aqualab 3TE water activity meter (Cicuttin and
(2009) was utilised, which is related with the shrinkage model of Giner, 2006). No net effect of temperature was found, in agreement
Eq. (19). The density model predicts: with Leiva Díaz et al. (2009). Moisture contents were measured in
atmospheric oven at 105 °C until constant weight (Greensmith,
q d ð1 þ W Þ
q¼ ð24Þ 1998). In this high initial moisture product, the water activity
1 þ qqd W
w interval was divided in two, leading to high and low-moisture Hal-
sey models (see Appendix).
2.2.4.3. Air-product heat and mass transfer coefficients. These were
measured in previous experiments (Cicuttin and Giner, 2006) of 3. Results and discussion
steady-state water evaporation carried out in the same
mechanical-convection oven at an air velocity of 0.2 m/s .The 3.1. Wheat drying
average heat transfer and mass transfer coefficient were found to
be hT = 21.9 W m2 K1 and kp = 1.51 107 kg water vapour/ 3.1.1. Variation of the mass transfer Biot number for high moisture
(m2 s Pa). Existing correlations differ considerably in the low air contents
velocity zone (Welti-Chanes et al., 2003; Marinos-Kouris and Mar- The exact analytical solution for Eq.(1), with initial and bound-
oulis, 1995), thus the need of the experimental determinations. ary conditions (2.1), (2.2), (2.3) (Crank, 1975), assumes that the
Biot number is constant, i.e., not time-dependent. Few if any con-
2.2.4.4. Variable thickness. The domain of Eq. (14), i.e., x from 0 to L tributions were provided by the literature to depict the variation
was discretised in I nodes, so I 1 distance intervals Dx were cre- of the mass transfer Biot number during drying. For this purpose,
ated, with x = (i 1) Dx. Then, the space index ‘‘i” varies from 1 to I the diffusion equation mentioned above was numerically solved
for nodes and from 1 to I 1 for intervals. The time, in turn, was by an implicit finite-difference scheme. The radius domain was
discretised as t = (nt 1) Dt. Therefore the leather thickness at discretised in 31 nodes, using a time step of 1 s. A Matlab program
any time during drying was calculated by: was written for this purpose, being the predictions completely
stable. The mass transfer Biot number was studied as a function
X
I1
of time for wheat drying, firstly from an initial moisture content
L¼ Dxi;nt ð25Þ of 0.35 kg water/kg dry matter. This unusually high water content
i¼1
for wheat was selected in order to clarify the predicted behaviour
Each space interval is reduced after each time interval using the plotted in Fig. 3.
shrinkage expression (Eq. (19)). Being the mass transfer coefficient, sphere radius, dry matter
density relationship and effective diffusion coefficient assumed
2.2.5. Calculation of the heat transfer Biot number during drying constant, the only variable factor is the derivative of moisture con-
Given the absence of published data for this type of leather, tent with respect to water activity (hereafter slope). This slope is
the thermal conductivity for apples in J m1 K1 was utilised placed in the denominator of Eq. (10), so the mass transfer Biot de-
(Mohsenin, 1980). pends reciprocally on it. The behaviour of oW/oaw is typical of a
S.A. Giner et al. / Journal of Food Engineering 101 (2010) 214–222 219
0.4
0.3
0.25
0.2
0.15
0.1
0.05
0
0 10 20 30 40 50 60
Tiempo, s
Fig. 4. Predicted surface moisture content as a function of time during drying, for
the conditions of Fig. 3. Surface equilibrium would be approached in some 35 s, and
strict internal control to the mass transfer rate would prevail from then on.
will be very low initially, will then increase and decrease again.
However, the decrease is mechanistically unimportant, since it oc-
curs well inside the zone of strict internal control. The variation of
the derivative would explain the predicted peak in the graph of the 3000
mass transfer Biot in Fig. 3 (top) as a function of time. The initial
predicted value of this Biot number under these conditions was
2.1, suggesting that control of the mass transfer rate is initially 2000
shared between the internal and the external resistances. How-
ever, this state of affairs lasts less than 20 s (Fig. 3 bottom) as inter-
nal control takes over from then on, i.e., the mass Biot increases
1000
dramatically with time. Another evidence of the sudden increase
of the internal resistance to mass transfer within the first 30 s is
provided by Fig. 4, where the surface moisture content is predicted
to drop very sharply from the initial moisture content, towards the 0
equilibrium value with the drying air conditions. 0 10 20 30 40 50
Time, s
3.1.2. Variation of the mass transfer Biot number at lower, practical
Fig. 5. Mass transfer Biot number as a function of time during drying of a wheat
moisture contents grain. Operating conditions: initial moisture content of wheat, 0.205 kg water/kg
At an initial moisture content W0 of about 0.205 (17% w/w), dry matter (17% w/w); air temperature, 60 °C; air relative humidity, 0.08; air mass
which is a typical water content of freshly harvested wheat fed flow rate, 0.30 kg m2 s1; and grain equivalent radius, 1.8.0 103 m.
220 S.A. Giner et al. / Journal of Food Engineering 101 (2010) 214–222
1
0.9
0.8
Heat Transfer Biot Number
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0 10 20 30 40 50 Fig. 7. Conventional (—) and modified (—) heat transfer Biot number predicted as a
function of time during the drying of low-calorie, apple leathers. The modified
1.2 number suggests a more externally controlled heat transfer rate during drying than
that predicted by the conventional definition. This emphasizes the mass transfer-
controlled nature of food drying.
1
Heat Transfer Biot Number
8.E-10
Effective diffusion coefficient, m 2/s
7.E-10
3.1.3. Evolution of the heat transfer Biot number in wheat drying
For an initial moisture content of 0.35 kg water/kg dry matter,
6.E-10
the variation in the heat transfer Biot number (Fig. 6 top) is gov-
erned by the variation of the surface thermal conductivity. The
5.E-10
lower the surface moisture content, the lower the thermal con-
ductivity and the higher the Biot number (see Eq. (12)): in the
4.E-10
first seconds, the surface moisture content in Fig. 4 decreases
only slightly and then the heat transfer Biot number remains
3.E-10
substantially constant. However, then the surface moisture
drops sharply to reach a constant value, and, simultaneously,
2.E-10
the Biot increases suddenly to reach a value of about 0.95 after
some 30 s.
1.E-10
Concerning wheat drying from an initial moisture content of
0.205 kg water/kg dry matter, the evolution of the heat transfer
0.E+00
Biot number is simpler. From an initial value of 0.6, it increases
0 1 2 3 4 5
to 0.95, to remain then constant due to attainment of equilibrium
conditions. The final value is the same for both drying conditions, Moisture content, kg water/kg dry matter
since it depends, through the thermal conductivity, on the surface
Fig. 8. Diffusion coefficient for low-calorie apple leathers as a function of moisture
moisture content, which tends to a value in equilibrium with the content for product temperatures of 30 (. . .), 50 (—) and 70 °C (—) as predicted with
same air conditions. the correlation developed for this shrinking food.
S.A. Giner et al. / Journal of Food Engineering 101 (2010) 214–222 221
1.5 4. Conclusions
an analytical solution of a diffusion equation with internal control Chen, X.D., Peng, X., 2005. Modified Biot number in the context of air drying of small
moist porous objects. Drying Technology 23, 83–103.
would be defective for the initial 35 min of drying and is therefore
Cicuttin, S.R., Giner, S.A., 2006. Geles pécticos semideshidratados dietéticos de
not recommended. manzana. Simulación numérica considerando contracción volumétrica
The heat transfer Biot number both in wheat and low-calorie (Semidehydrated low-calorie apple pectic gels. Numerical simulation
apple-leather keeps within the internal–external control region considering shrinkage). In Anais do Simposio Brasileiro Sobre
Desenvolvimento de Novos Produtos Alimenticios. Alimentos Funcionais e
during the drying process. para Fins Especias. Campinas-SP, Brazil.
A modified version of the heat transfer Biot for drying tends Crank, J., 1975. The Mathematics of Diffusion, second ed. Oxford University Press,
asymptotically to zero during the process because the heat influx Oxford, UK.
Crapiste, G.H., Rotstein, E., 1997. Design and performance evaluation of dryers. In:
to the product is lower than in pure heat transfer due to the frac- Valentas, K.J., Rotstein, E., Singh, R.P. (Eds.), Handbook of Food Engineering
tion of heat demanded for water evaporation at the surface. Practice. CRC Press, Boca Raton, pp. 125–165.
This is congruent with the predictions of practically flat internal Gastón, A., Abalone, R., Giner, S.A., Bruce, D.M., 2004. Effect of modelling
assumptions on the effective water diffusivity in wheat. Biosystems
temperature profile during drying of foods. The modified heat Engineering 88 (2), 175–185.
transfer Biot during drying. Gely, M.C., Giner, S.A., 2007. Diffusion coefficient relationships during drying of soya
As a consequence of the above conclusions, numerical solutions bean cultivars. Biosystems Engineering 96 (2), 213–222.
Giner, S.A., Denisienia, E., 1996. Pressure drop through wheat as affected by air
considering shrinkage and variable diffusion coefficient are sug- velocity, moisture content and fines. Journal of Agricultural Engineering
gested for drying kinetic studies in high moisture foods. Research 63, 73–86.
Giner, S.A., Mascheroni, R.H., 2001. Diffusive drying kinetics in wheat. Part 1:
potential for a simplified analytical solution. Journal of Agricultural Engineering
Appendix. Research 80 (4), 351–364.
Giner, S.A., Mascheroni, R.H., 2002. Diffusive drying kinetics in wheat. Part 2.
Applying the simplified analytical solution to experimental data. Biosystems
The Halsey equations are as follows: Engineering 81 (1), 85–97.
1 Giner, S.A., 2009. Influence of internal and external resistances to mass transfer on
Expð2:502Þ 0:9017 the constant drying rate period in high moisture foods. Biosystems Engineering
W¼ ðA1Þ 102, 90–94.
Lnaw Greensmith, M., 1998. Practical Dehydration, second ed. Woodhead Publishing Ltd,
Cambridge, England.
fitted with r2 = 0.996, and valid for 0.855 6 aw 6 0.999 while Gurney, H.P., Lurie, J., 1923. Charts for estimating temperature distributions in
1 heating and cooling solid shapes. Industrial and Engineering Chemistry 15 (11),
Expð2:367Þ 1:100 1170–1172.
W¼ ðA2Þ Kechaou, N., Roques, M.A., Lambert, J.F., 1987. Diffusion in shrinking media: the case
Lnaw of drying of gels. In: Jowitt, R. (Ed.), Physical Properties of Foods, vol. 2. Elsevier
Applied Science, London, UK, pp. 55–60.
fitted with r2 = 0.998 and valid for 0.1 6 aw < 0.855. Leiva Díaz, E., Giannuzzi, L., Giner, S.A., 2009. Apple pectic gel produced by
dehydration. Food and Bioprocess Technology 2 (2), 194–207.
Kaymak-Ertekin, F., Gedik, A., 2004. Sorption isotherms and isosteric heat of
References sorption for grapes, apricots, apples and potatoes. LWT Food Science and
Technology 37, 429–438.
Aguilera, J.M., 2005. Why food microstructure? Journal of Food Engineering 67, 3– Marinos-Kouris, D., Maroulis, Z.B., 1995. Transport properties in the drying of solids.
11. In: Mujumdar, A.S. (Ed.), Handbook of Industrial Drying, second ed., vol. 1.
Akpinar, E., Midilli, A., Bicer, Y., 2003. Single layer behaviour of potato slices in a Marcel Dekker, Inc., New York.
convective cyclone dryer and mathematical modelling. Energy Conversion and Mohsenin, N.M., 1980. Thermal Properties of Food and Agricultural Materials.
Management 44, 1689–1705. Gordon and Breach, London, UK.
Adhikari, B., Howes, T., Lecomte, D., Bhandari, B.R., 2005. A glass transition Parti, M., 1993. Selection of mathematical models for drying grain in thin layers.
temperature approach for the prediction of the surface stickiness of a drying Journal of Agricultural Engineering Research 54, 339–352.
droplet during spray drying. Powder Technology 149, 168–179. Singh, R.P., Heldman, D.R., 1993. Introduction to Food Engineering, second ed.
Brooker, D.B., Bakker-Arkema, F.W., Hall, C.W., 1992. Drying and Storage of grains Academic Press, New York.
and oilseeds. Van Nostrand Reinhold, New York. Sokhansanj, S., Bruce, D.M., 1987. A conduction model to predict grain temperatures
Bruce, D.M., 1985. Exposed-layer barley drying: three models fitted to new data up in grain drying simulation. Transactions of the ASAE 30 (4), 1181–1184.
to 150 °C. Journal of Agricultural Engineering Research 32, 337–347. Welti-Chanes, J., Mujica-Paz, H., Valdez-Fragoso, A., Leon-Cruz, R., 2003.
Beuchat, L.R., 1981. Microbial stability as affected by water activity. Cereal Foods Fundamentals of mass transport. In: Welti-Chanes, J., Vélez-Ruiz, J., Barbosa-
World 26 (7), 345–349. Cánovas, G.V. (Eds.), Transport Phenomena in Food Processing. CRC Press LLC,
Beuchat, L.R., 1983. Influence of water activity on growth, metabolic activities and Boca Raton, USA.
survival of yeasts and molds. Journal of Food Protection 46, 135–141.